作为eSTREAM计划的7个最终获选算法之一,Sosemanuk至今未见有效的攻击算法.针对流密码Sosemanuk的线性弱点,利用线性掩码技术比特异或代替模加和Trans函数运算,把非线性函数转化为线性函数,依据线性逼近建立区分器;设计了一种基于Sosemanuk中有限状态机(粥M)和Serpentl线性特征的区分攻击.结果表明,需要2221bit就能将流密码Sosemanuk的密钥流序列从随机序列中区分出来.
Sosemanuk has successfully been chosen as one of the final stream ciphers by eSTREAM. Although Sosemanuk has attracted great interest, it has not been completely developed. Due to the linear defect in the Sosemanuk stream cipher, a distinguishing attack based on the linear character of the finite state machine (FSM) and Serpentl was proposed. This method used linear masks by replacing modular additions and Trans functions by exclusive ORs (XORs), transforming the nonlinear functions into linear functions. Then, a distinguisher was built. The results show that the keystream generated in Sosemanuk is distinguishable from a random sequence after observing approxi- mately 2^221 bits.